Debugging Regex

The first move: paste it into regex101

Whatever the bug, your first action should be: copy the pattern, copy the input, paste both into regex101.com with the right flavor selected. The explanation pane decomposes the pattern; the match panel shows what hit and what didn't.

Half of regex bugs are invisible until you see them visualized.

Debug flag in Python

re.compile(pattern, re.DEBUG) prints the compiled state machine. Useful for understanding how the engine actually parses your pattern.

The narrowing technique

When a pattern "doesn't match what it should," cut it in half:

Take the first half of the pattern, test against the input. Does it match?
Take the second half. Does the input have what it expects?
Find the boundary where the failure starts.

This bisection finds the issue faster than staring at the whole pattern.

Logging captures

If the issue is "matches but captures wrong stuff," print every group:

m = pattern.search(text)
if m:
    print(m.group(0))      # whole match
    print(m.groups())      # numbered groups
    print(m.groupdict())   # named groups
    print(m.span())        # position

Common bug signatures

"Returns None": probably anchoring or wrong escape (raw string?).
"Matches too much": greedy vs lazy, or missing word boundaries.
"Matches too little": missing optional pieces, wrong character class.
"Hangs": catastrophic backtracking. Look for nested quantifiers.
"Works in tester, fails in code": flavor mismatch or string escape issue.

Code

Debugging tools·python

import re

# Print the compiled state machine
re.compile(r'(a|ab)c', re.DEBUG)
# Outputs the engine's view of the pattern

# Walk the input position by position
pattern = re.compile(r'(\w+)\s+(\w+)')
text = 'hello world foo bar'
for m in pattern.finditer(text):
    print(f'pos {m.start():3} matched {m.group()!r} groups={m.groups()}')

# Quick verification: does the pattern even compile cleanly?
try:
    re.compile(r'your pattern')
except re.error as e:
    print(f'pattern syntax error: {e}')

# Test multiple inputs at once
for inp in ['known good', 'known bad', 'edge case']:
    m = pattern.search(inp)
    print(f'{inp!r}: {m.group() if m else None}')

The first move: paste it into regex101

Debug flag in Python

The narrowing technique

Logging captures

Common bug signatures

Code

External links

Exercise

Progress

Comments 0